Cr - ACS Publications

Sep 27, 2017 - Laboratory of Industrial Chemistry, Ruhr-University Bochum, 44780 Bochum, North Rhine-Westphalia, Germany. ‡. Tyndall National Instit...
0 downloads 0 Views 1MB Size
Subscriber access provided by LONDON METROPOLITAN UNIV

Article

Spinel-Structured ZnCr2O4 with Excess Zn is the Active ZnO/ Cr2O3 Catalyst for High-Temperature Methanol Synthesis Huiqing Song, Daniel Laudenschleger, John J Carey, Holger Ruland, Michael Nolan, and Martin Muhler ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b01822 • Publication Date (Web): 27 Sep 2017 Downloaded from http://pubs.acs.org on September 29, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Spinel-Structured ZnCr2O4 with Excess Zn is the Active ZnO/Cr2O3 Catalyst for High-Temperature Methanol Synthesis Huiqing Song†, Daniel Laudenschleger†, John J. Carey‡, Holger Ruland†, Michael Nolan‡*, and Martin Muhler†* †

Laboratory of Industrial Chemistry, Ruhr-University Bochum, 44780 Bochum, North Rhine-Westphalia, Germany ‡Tyndall National Institute, University College Cork, Cork T12R5CP, Munster, Ireland

ABSTRACT: A series of ZnO/Cr2O3 catalysts with different Zn:Cr ratios was prepared by co-precipitation at a constant pH of 7 and applied in methanol synthesis at 260-300 °C and 60 bar. The X-ray diffraction (XRD) results showed that the calcined catalysts with ratios from 65:35 to 55:45 consist of ZnCr2O4 spinel with low degree of crystallinity. For catalysts with Zn:Cr ratios smaller than 1, the formation of chromates was observed in agreement with temperature-programmed reduction results. Raman and XRD results did not provide evidence for the presence of segregated ZnO indicating the existence of Zn-rich non-stoichiometric Zn-Cr spinel in the calcined catalyst. The catalyst with Zn:Cr=65:35 exhibits the best performance in methanol synthesis. The Zn:Cr ratio of this catalyst corresponds to that of the Zn4Cr2(OH)12CO3 precursor with hydrotalcite-like structure obtained by co-precipitation, which is converted during calcination into a non-stoichiometric Zn-Cr spinel with an optimum amount of oxygen vacancies resulting in high activity in methanol synthesis. Density functional theory calculations are used to examine the formation of oxygen vacancies and to measure the reducibility of the methanol synthesis catalysts. Doping Cr into bulk and the (10-10) surface of ZnO does not enhance the reducibility of ZnO, confirming that Cr:ZnO cannot be the active phase. The (100) surface of the ZnCr2O4 spinel has a favorable oxygen vacancy formation energy of 1.58 eV. Doping this surface with excess Zn charge-balanced by oxygen vacancies to give a 60% Zn content yields a catalyst composed of an amorphous ZnO layer supported on the spinel with high reducibility confirming this as the active phase for the methanol synthesis catalyst. KEYWORDS: ZnO/Cr2O3, methanol synthesis, non-stoichiometric spinel, oxygen vacancy, DFT+U

1. Introduction Methanol is one of the most important industrially produced basic chemicals due to its application in a wide range of fields.1-2 The investigation of methanol synthesis from a mixture of hydrogen and carbon oxides using metals and metal oxides as catalyst dates back to the 1910s.3 In the 1920s BASF issued two patents focusing on ZnO/Cr2O3 and Cu/ZnO catalysts, which were primarily used in the early period of the industrial production of methanol.4-5 Later, the Cu/ZnO/Al2O3 catalyst with high activity and selectivity was patented by ICI in the 1960s, which requires highly purified synthesis gas.6-7 Due to the lower operating temperature and pressure, the Cu/ZnO-based catalyst has been used extensively for methanol production in the last decades. However, Zn-Cr mixed oxides, which are applied in the synthesis of methanol at high temperatures and pressures,8 remain of interest owing to their resistance against sulfur impurities in syngas as well as the possibility of modifying their selectivities towards the synthesis of higher alcohols.9-12 Furthermore, due to the high thermal stability, ZnO/Cr2O3 catalysts demonstrated great potential in combination with microporous acidic catalysts in the one-pot synthesis of liquid hydrocarbons and short chain olefins from syngas.13-14 In the early investigation on methanol synthesis over ZnO/Cr2O3 catalysts prepared by co-precipitation followed by

calcination, the synergy observed between ZnO and Cr2O3 in comparison to the pure component oxides was mainly ascribed to an increase in the specific surface area.15 In addition, ZnO had been generally identified as the only active component, and the promoting role of Cr2O3 was attributed to spinel formation, which prevented the sintering of small ZnO crystallites by providing a considerably larger surface area and better distribution of ZnO particles.8 It is generally accepted that methanol synthesis over ZnO is a structure-sensitive reaction, in which oxygen vacancies on ZnO (000-1) are assumed to be the active sites. One model, which was reported by Boccuzzi et al.16 in 1978, described an active site composed of three Zn ions around an oxygen vacancy. Two years later, an active site with similar geometry and composition was proposed by Kung,17 who assumed that oxygen vacancies could assist in the adsorption and activation of CO. He proposed a catalytic cycle of methanol formation based on the oxygen vacancy as active site. The reaction starts with the dissociative adsorption of H2 on ZnO sites around the oxygen vacancy followed by the adsorption of the CO molecule in the oxygen vacancy. The reaction proceeds through hydrogenation of adsorbed CO by the transfer of adsorbed hydrogen to form surface formyl species (CHO)ads. Further hydrogenation of the formyl species yields adsorbed formaldehyde (H2CO)ads, which is hydrogenated forming methoxy species (H3CO)ads. The hydrogenation 1

ACS Paragon Plus Environment

ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of methoxy species leads to the product methanol. It was proposed that a change of the mechanism occurs when CO2 is present, as the very stable formate (HCO2)ads species may be formed in the defect site. This mechanism was confirmed by Kurtz et al.,18 who found that in the presence of CO2 the rate of methanol formation is significantly slower. In addition to ZnO, non-stoichiometric Zn-Cr spinel was reported as the active phase in methanol synthesis. Del Piero et al.19 first reported that the non-stoichiometric Zn-Cr spinel was formed by the dissolution of excess ZnO in ZnCr2O4, which results from the solid reaction between ZnO and Cr2O3 during calcination. They also pointed out that the dissolved Zn2+ ions may be located in octahedral sites randomly substituting Cr3+ ions, leaving some of the tetrahedral sites vacant. The deactivation of the catalyst was observed to be accompanied by a decrease in the non-stoichiometry of the spinel-like phase and a corresponding increase in the amount of crystalline ZnO detected by X-ray diffraction (XRD). Therefore, they assumed the non-stoichiometric spinel-like phase to be the active phase in the ZnO/Cr2O3 catalyst. More studies regarding the influence of the non-stoichiometry of the Zn-Cr spinel on its bulk and surface properties were carried out by Bertoldi et al.,20 who proposed a general chemical formula of the nonstoichiometric Zn-Cr spinel as ZnxCr2/3(1-x)O, where x ranges from 0.25 to 0.40. They also found that the activity of the nonstoichiometric Zn-Cr spinel towards CO adsorption is much higher than observed for pure ZnO, Cr2O3 and ZnCr2O4. Giamello et al.21 and Riva et al.22 also confirmed the formation of the non-stoichiometric Zn-Cr spinel based on methanol temperature-programmed desorption (TPD) and X-ray photoelectron spectroscopy (XPS) studies. Grimes et al.23 used atomistic simulation to explain the mechanisms of the dissolution of ZnO and suggested that one oxygen vacancy was created for every three ZnO species dissolved. The Zn:Cr ratio was found to significantly influence the catalytic activity of the ZnO/Cr2O3 catalyst.24 The optimum Zn:Cr ratio has been extensively investigated in the literature.8, 25-26 Molstad and Dodge25 reported that catalysts with Zn:Cr ratios between 70:30 and 60:40 exhibit the highest yield of methanol at relatively low temperatures of 300-325°C. This observation is in good agreement with the results reported by Bone,24 who found that the maximum activity is achieved with a Zn-excess catalyst containing about 75 atom% of Zn. According to Molstad and Dodge,25 the highest activity was obtained with the catalyst containing 20-30 atom% of Cr at a reaction temperature in the range from 350 °C to 425 °C. Additionally, they observed that the catalysts with Zn:Cr > 75:25 lose their activity rapidly, while the catalysts containing more Cr exhibit a slight activity increase with reaction time. Furthermore, after testing the catalysts had undergone volume shrinkage, which increased with increasing Cr content. When using the already shrunk catalyst, the maximum yield of methanol was obtained with a Zn:Cr ratio of around 1:1. These results were also confirmed by Errani et al.,26 however, with a different explanation. Based on the investigation of non-stoichiometric Zn-Cr spinel,19-22 the authors suggested that the activity increases up to Zn:Cr = 50:50 due to the increasing non-stoichiometry of the spinel-type phase, while the formation of a segregated ZnO phase causes a decrease in activity. Nevertheless, no linear correlation was found between the catalytic activity and the excess zinc inside the non-stoichiometric spinel-type structure up to Zn:Cr = 50 : 50. More recently, Bradford et al.27 reported that ZnO/Cr2O3 catalysts prepared by co-precipitation exhibit-

Page 2 of 15

ed an maximum specific activity for Zn:Cr = 2.56 (72:28), which is in good agreement with Molstad and Dodge.25 Although the influence of the Zn:Cr ratio on the performance of ZnO/Cr2O3 catalyst has been investigated over decades, there is still no agreement on the optimal Zn:Cr ratio. In this work, we studied the role of the Zn:Cr ratio in the catalytic activity of ZnO/Cr2O3 catalysts by a combined experimental and density functional theory (DFT) approach. A series of ZnO/Cr2O3 catalysts with different Zn:Cr ratios varying from 100:0 to 0:100 were prepared by co-precipitation at 65°C and pH 7 with Na2CO3 solution as the precipitating agent. The resulting precipitates were calcined in air at 320 °C for 3 h and characterized by N2 physisorption, XRD, temperatureprogrammed reduction (TPR), XPS and Raman spectroscopy. The catalytic testing was performed at 60 bar using a syngas mixture with H2:CO = 1.5, and the reaction temperature was varied in steps of 20 °C from 260 to 300 °C. The catalyst with Zn:Cr=65:35 exhibits the best performance in methanol synthesis. Given that previous work shows the importance of oxygen vacancies in methanol synthesis, we combine these experiments with DFT calculations of the reducibility of different Cr/Zn-containing systems, as measured by the oxygen vacancy formation energy. For a catalyst composed of Cr doping into ZnO (as excess Zn), the oxygen vacancy formation energies are larger than undoped ZnO. The oxygen vacancy formation energies for the pure ZnCr2O4 spinel are significantly lower than the undoped ZnO. Doping the spinel structure with excess Zn on Cr3+ octahedral sites gives compositions ranging from 30% to 60% Zn content. At the higher concentration of incorporated Zn, the Zn dopants replace all surface Cr atoms (it is less favorable to replace bulk Cr atoms) and the oxygen vacancy formation energy decreases indicating high reducibility of this system. The structure of this catalyst can be thought of as a thin, amorphous ZnO-stoichiometry layer supported on the ZnCr2O4 spinel. In addition, the adsorption of CO and H2 on ZnCr2O4 with excess Zn is weaker than on stoichiometric ZnCr2O4, allowing for increased reactivity. These DFT insights help to explain the experimental observations that both ZnO and ZnCr2O4 are present in the catalyst. The reducibility of and interaction with CO and H2 on ZnCr2O4 are the key to increased methanol activity and a superior methanol synthesis catalyst. 2. Experimental a. Catalyst preparation Catalysts with different Zn:Cr ratios were prepared by a coprecipitation procedure adapted from Kurtz et al.18 A metal salt solution (1 M) consisting of zinc and chromium nitrates with the desired Zn:Cr ratio was added dropwise into a vessel, which contained 50 ml HPLC water heated up to 65°C.The precipitation was performed under continuously stirring at a constant pH of 7 by simultaneous addition of Na2CO3 solution (1.2 M) using an autotitrator while the precipitation temperature was constantly kept at 65°C. The suspension after precipitation was aged at 65 °C for 2 h. Subsequently, the precipitate was filtered and washed with HPLC water and then dried for 18 h at 120 °C. The dried precipitate was granulated and calcined in synthetic air for 3 h with a heating rate of 2 °C min−1 to 320 °C and keeping this temperature for 3 h. b. Characterization 2

ACS Paragon Plus Environment

Page 3 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

The metal composition as well as the content of residual Na were determined by optical emission spectroscopy with inductively coupled plasma (ICP-OES) using an UNICAM PU 7000. Powder XRD was performed using a Panalytical MPD diffractometer with Cu Kα radiation (λ = 0.154 nm) at 45 kV and 40 mA without monochromator. A Ni filter was used to suppress the Kß emission line.The XRD patterns were recorded over a 2θ range of 5–85° with a 0.0131 step width and 250 s collection time. Powder diffraction files (PDF) from the International Center of Diffraction Data (ICDD) and X’PertHighScore Plus software were applied for qualitative phase analysis. Specific surface areas and pore structures were measured by N2 physisorption at 77 K using a Quantachrome Autosorb-1-MP instrument. The catalysts were degassed under vacuum at 200 °C for 2 h prior to the measurement. Pore volume and average pore diameter were obtained using the Barrett-Joyner-Halenda (BJH) method. TPR was carried out in a U-tube quartz reactor with a H2/Ar mixture containing 4.6% H2 as reducing agent. About 100 mg catalysts (250–355 µm sieve fraction) were flushed with Ar (99.9995%) to remove weakly bound water and then reduced using a total flow rate of 84.1 Nml/min at a rate of 5 °C/min with a programmable temperature controller. The samples were heated up to 800 °C and temperature was held at 800 °C for 1 h. H2 consumption was recorded by a thermal conductivity detector (TCD). The effluent gas passed through a cold trap filled with isopropanol and dry ice before entering the TCD in order to remove water formed during reduction. XPS measurements were carried out in an ultra-high vacuum set-up equipped with a highresolution Gammadata-Scienta SES 2002 analyzer and a monochromatic Al Kα as the X-ray source (1486.6 eV, anode operating at 14 kV and 30 mA). The base pressure in the measurement chamber was maintained at about 7 x 10-10 mbar. The measurements were performed in the fixed transmission mode with a pass energy of 200 eV. Charging effects during the measurements were compensated by applying a flood gun. The measured spectra were calibrated based on the C 1s peak at 284.5 eV originating from adsorbed hydrocarbons. The CasaXPS software was used to analyze the XPS data. Raman spectra were measured using a WITec confocal Raman microscope Alpha300 R/S/A. The samples excitation was carried out by using a 532 nm frequency-doubled Nd:YAG laser. A 50x/0.4 NA objective was used to focus the laser onto the sample. The back-scattered Raman light was collected by the same objective and transferred via a 50 µm multimode fiber to the spectrometer unit consisting of a 600 lines/mm diffraction grating and a 1600x200 pixels electron multiplying charge-coupled device camera operating at 60 °C. The spectral resolution was approximately 4 cm-1. 10 Raman spectra of each sample were recorded with an integration time of 30 s for each spectrum at a laser power of 20 mW. c. Continuous methanol synthesis set-up The set-up for methanol synthesis from syngas (Figure S1) consists of the gas supply unit, the reactor unit and the gas chromatograph as analytical device. The gas supply includes syngas, Ar, H2, and H2 diluted in He. In the present work, syngas with a ratio of H2:CO =1.5 containing 10% N2 as internal standard was used for the catalytic tests. In the reactor unit the main gas line was split into 4 branch lines connected to 4 separate reactors. Each branch line consisted of a MFC, a valve, a back-pressure valve and two filters. All 4 reactors were mounted in an oven, which established the

reaction temperature in the range from 260 °C to 300 °C. The desired pressure for the reaction was controlled by a pressure regulator with proportional-integral-differential (PID) controllers. The outlet gas of the chosen reactor was introduced into the GC by switching the multi-position valve (VM), while the gas mixtures from the other reactors were passed to the exhaust. In order to avoid product condensation in the gas mixture, all gas lines from the reactor outlet to the GC were installed in a separate oven, in which the temperature was set to 155 °C. The reactors were made of stainless steel and coated with glass on the inner wall. The catalyst was fixed between two plugs of quartz wool in the middle of the reactor, and two glass rods were placed in the reactors to reduce the dead volume. d. Catalytic tests 500 mg of catalyst with particle sizes of 250–355 µm was loaded in each glass-lined stainless steel reactor. No prereduction was carried out prior to the catalytic tests. Methanol synthesis was performed at 60 bar using a mixture of 36% CO, 54% H2 and 10% N2 at a volumetric feed rate of 40 ml min−1 (STP) resulting in a GHSV of 4800 ml h−1 gcat−1. The reaction temperature was varied in steps of 20 °C from 260 to 300 °C. At each temperature the reaction was carried out for 20 h. Online gas analysis was performed with an Agilent gas chromatograph equipped with two thermal conductivity detectors using He as carrier gas. e. Computational Methodology All calculations are carried out using density functional theory (DFT) with the generalized gradient approximation (GGA)28 and the Perdew-Burke-Ernzerhof (PBE)29 exchange correlation functional as implemented in the Vienna ab initio Simulation Package (VASP).30-31 The valence electrons are expanded using a plane wave basis set, and the interactions between the core (Cr:[Ar], Zn:[Ar], O:[He]) and the valence electrons (Cr:[3p6, 3d5, 4s1], Zn:[3d10, 4s2], O:[2s2, 2p4]) are treated using the projected augmented wave (PAW) method.32 To describe the strong electron correlation of Cr cations in CrZnO and in the spinel structure, a Hubbard +U correction (giving DFT+U)33-34 is applied to the Cr 3d states, for which the value of U is 5 eV.35-36 Due to doping of lower valent Zn2+ cations onto Cr3+ sites, oxygen hole states will form and in describing these states we apply U=5.5 eV36 to the O 2p states. The low energy structures for the ZnO bulk and ZnCr2O4 spinel bulk are obtained by relaxing a series of structures from -4% to +4% of the experiment lattice constant where the atomic positions, cell vectors and angles were allowed to relax at a constant volume. The energies obtained from each structure were fitted to the Murnaghan Equation of State37 to obtain the minimum energy structure. This was carried out for different plane wave cut-off energies of 400 eV, 450 eV and 500 eV where the Brillouin Zone integrations were sampled using the Monkhorst-Pack method38 at different k-point sampling grids of 2x2x2, 4x4x4, and 6x6x6. The structures were deemed converged when the forces on the atoms were less than 0.02 eV/Å. For both bulk structures, the parameters for the bulk structure are a plane wave cut-off energy of 450 eV and a kpoint sampling grid of 4x4x4 which gives a deviation of 1.6% from the experimental lattice constants for both bulk structures.39-40 The ZnO bulk structure was expanded to a 3x3x3 supercell (Zn54O54) and relaxed using a reduced Gamma centered kpoint sampling grid of 1x1x1 with the same energy cut-off, 3

ACS Paragon Plus Environment

ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

while the ZnCr2O4 spinel bulk was expanded to a 1x1x2 supercell structure (Zn16Cr32O64) with a reduced k-point sampling grid of 4x4x2. The bulk structures were expanded to accommodate point defects under periodic boundary conditions for investigations of Cr doping and oxygen vacancy formation in each of the systems of interest. The non-polar ZnO (10-10) and ZnCr2O4 (100) surfaces are also examined as model catalysts to explore different possible structures and compositions, with Cr:Zn ratios ranging from 0 – 100%. The surfaces are generated from the lowest energy bulk structures and modelled using the slab method. For ZnO (10-10) and ZnCr2O4 (100), slab thicknesses of 13 Å and 10 Å, with 9 and 11 atomic layers, are used. The vacuum region above the surface is 13 Å. The bottom 3 layers of each slab are fixed while the rest of the slab is allowed to fully relax. The surfaces are converged using the same parameters as the bulk calculations The different compositions of the Cr:ZnO systems in the experimental catalysts are modelled by Cr3+ doping in ZnO and Zn2+ doping on a Cr site in ZnCr2O4 spinel using the bulk and surface structures. A key property identified experimentally for methanol synthesis catalyst is the reducibility, which is modelled by oxygen vacancy formation. We investigate the effect of doping on the reducibility. The doping of these structures is carried out in a similar manner to previous studies for charge compensating mechanisms for doping in Cr2O3,36 CeO2,41-43 or TiO244 by first examining any charge compensating oxygen vacancies in the doped system and then calculating the formation energy of the next oxygen vacancy, which reduces the oxide. We also investigate the interaction with CO and H2 by adsorbing the molecules on different sites of the stoichiometric ZnCr2O4 and Zn-rich ZnCr2O4 (100) surfaces, computing the adsorption energy of each molecule as follows: Eads = E(CO/H2-ZnCr2O4) – [E(CO/H2) + E(ZnCr2O4)] Where E(CO/H2-ZnCr2O4) is the total energy of CO or H2 adsorbed at the spinel (100) surface, E(CO/H2) is the total

Page 4 of 15

energy of the free molecule and E(ZnCr2O4) is the total energy of the spinel (100) surface. The changes in oxidation states of the cations in the bulk structures was investigated using Bader’s atoms in molecules method as implemented in VASP by Henkelman et al.45-47 and spin magnetization values. 3. Results a. Characterization of the catalyst precursors The actual metal compositions of the calcined precursors determined by ICP-OES as well as the textural properties determined by N2 physisorption are summarized in Table 1. Small variations between the nominal and actual Zn:Cr ratio can be observed for all the catalysts. More specifically, the actual amount of Cr in the catalysts is always less than expected. Based on the calculation of the maximal concentration of ionic Cr after precipitation, the possibility of incomplete precipitation of Cr3+ can be excluded. Hence, the loss of Cr probably occurred during the filtration. The Cr-containing precipitates formed with a very small particle size during the co-precipitation were most likely Cr(OH)3 nanoparticles,48 and a part of these precipitates penetrated the glass fiber filter during the filtration. For the catalysts containing high amounts of Cr, a considerably high amount of Na residue was found to be present. This difficulty of removing Na may be also caused by the small-sized Cr(OH)3 nanoparticles formed by coprecipitation. The reaction between acidic Cr(OH)3 and the basic Na residue presumably prevents the complete removal of Na.49 For the catalysts with lower Zn:Cr ratio, more smallsized Cr(OH)3 was formed, resulting in more difficult filtration and washing. The specific surface area strongly depends on the Zn:Cr ratio of the catalysts. With decreasing nominal Zn:Cr ratio from 100:0 to 60:40 the specific surface area increases, and for Zn:Cr= 60:40 the specific surface area reaches a maximum of 117 m2 g-1. A further decrease of the nominal Zn:Cr ratio from 60:40 to 0:100 results in a decrease in specific surface area.

Table 1. Nominal and actual metal composition as well as the textural properties of the catalysts after calcination. Catalyst

Nominal Zn:Cr ratio (atomic)

Actual Zn:Cr ratio (atomic)

Na (wt %)

ABET (m2 g-1)

rp (nm)

Vp (cm3 g-1)

100Zn

100:0

100:0

0.1

34

6.8

0.14

80Zn20Cr

80:20

83:17

0.2

82

2.4

0.11

70Zn30Cr

70:30

71:29

0.03

95

2.9

0.18

65Zn35Cr

65:35

67:33

0.04

106

2.7

0.18

60Zn40Cr

60:40

62:38

0.01

117

1.8

0.17

55Zn45Cr

55:45

57:43

0.3

112

2.4

0.18

50Zn50Cr

50:50

51:49

0.7

85

1.9

0.12

40Zn60Cr

40:60

46:54

2.8

34

1.9

0.07

33Zn66Cr

33:66

35:65

4.1

18

1.9

0.05

20Zn80Cr

20:80

28:72

7.2

2

2.4

0.00

100Cr

0:100

0:100

8.1

3

9.2

0.03

4

ACS Paragon Plus Environment

Page 5 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

When the composition of the catalyst is changed from excess of Zn to excess of Cr, the specific surface area undergoes a drastic decline. In this work, the catalysts containing high levels of Cr generally exhibit lower surface areas and smaller pore volumes compared with the catalysts containing high amounts of Zn due to the high contents of Na residue. This trend in the specific surface area of the catalysts scales with the catalytic activity shown in section b. Catalytic tests, indicating that surface area is a major factor that influences the catalytic activity of the ZnO/Cr2O3 catalysts. It should also be noted that both pure ZnO and Cr2O3 exhibit a significantly smaller surface area than the ZnO/Cr2O3 catalysts with high Zn:Cr ratio between 80:20 and 50:50. Therefore, in the ZnO/Cr2O3 catalyst system, Cr2O3 probably does not act as a traditional support material that directly increases the surface of the catalyst. Instead, ZnCr2O4 is the actual support that increases the surface area of the catalyst and improves the distribution of ZnO.8 The XRD patterns of the hydroxycarbonate precursors before calcination are shown in Figure 1. For 100Zn, most of the reflections can be assigned to zinc hydroxycarbonate. A few unidentified reflections possibly originate from Zn(OH)2 or/and ZnCO3. For the precursors containing both Cr and Zn, the existence of the Zn-Cr hydrotalcite-like phase Zn4Cr2(OH)12CO3 can be confirmed by identification of the corresponding characteristic reflections located at 2θ = 34.5° and 2θ = 61.2°.50-51 With decreasing nominal Zn:Cr ratio from 80:20 to 20:80, these two reflections become broader and their intensity decreases, which indicates that the precursors become more amorphous as more Cr is present. The XRD pattern of 100Cr shows no identifiable reflections at all. With decreasing Zn:Cr ratio, more Cr(OH)3 can be formed during the co-precipitation, which results in a more amorphous state of the precursors. It has to be pointed out that Cr2(CO3)3 is not the product of the precipitation of Cr3+ using Na2CO3 solution.52

70:30, the reflections of ZnO generally become broader and weaker, and a further decrease in the Zn:Cr ratio leads to the disappearance of the ZnO reflections. The ZnCr2O4 spinel phase is found to exist in the catalysts with the Zn:Cr ratios from 70:30 to 20:80. The catalysts containing high levels of Cr show sharper reflections of ZnCr2O4 compared with the catalysts containing more Zn. Furthermore, the sharp reflections of Cr2O3 can only be found in the XRD pattern of 100Cr. Additionally, a series of small and sharp reflections located between 2θ = 14° and 32° can be observed in the catalysts containing a high amount of Cr, which can be mostly assigned to zinc and sodium chromate. The presence of sodium chromates in the calcined catalyst corresponds to the results obtained by ICP-OES, which show a certain amount of Na residue to be present in the catalyst containing excess of Cr. Obviously, the residual Na compounds such as Na2CO3 react with Cr2O3 and O2 forming sodium chromates. In addition to the formation of sodium chromates, Cr2O3 can also be oxidized yielding ZnCrO4.

Figure 2. XRD patterns of the calcined ZnO/Cr2O3 catalysts.

Figure 1. XRD patterns of the hydroxycarbonate precursors.

Figure 2 shows the XRD patterns of the calcined ZnO/Cr2O3 catalysts. It can be seen that the hydroxycarbonates with hydrotalcite-like structure were decomposed by calcination at 320 °C for 3 h. The XRD pattern of 100Zn exhibits clear and sharp reflections of ZnO indicating the formation of relatively large ZnO particles. With decreasing Zn:Cr ratio from 100:0 to

The catalysts with a Zn:Cr ratio in the range from 55:45 to 33:66 only exhibit the broad reflections of the ZnCr2O4 phase in addition to the chromates phases. According to Table 1, a considerable amount of ZnO should be present in these catalysts due to the 1:2 ratio of Zn:Cr in ZnCr2O4, even if all Cr2O3 reacted to form ZnCr2O4 during the calcination, but no characteristic ZnO reflections were found. The absence of ZnO in the ZnO/Cr2O3 catalyst with low Zn:Cr ratio is in agreement with the results reported by Del Piero et al.,19 who suggested that ZnO can be dissolved in the ZnCr2O4 spinel to form a nonstoichiometric Zn-Cr spinel. Hence, a small shift of the ZnCr2O4 reflections is supposed to be detected. The quantity of non-stoichiometric Zn-Cr spinel can be estimated by determination of the lattice parameter of the spinel phase using quantitative XRD analysis.19,26 However, due to the broad reflections, this method cannot be applied in the present work. Figure 3 shows the effect of the Zn: Cr ratio on the formation of chromates and spinels during calcination more clearly. After calcination at 250 °C and 320 °C, the 65Zn catalyst exhibits a series of broad reflections indicating the existence of amorphous material. A further increase in temperature led to the appearance of the sharp reflections of ZnO and ZnCr2O4. Based on these observation and the XRD patterns of the used catalyst, it can be assumed that the catalyst contains 5

ACS Paragon Plus Environment

ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

amorphous ZnO and ZnCr2O4. The 33Zn catalyst shows sharp reflections of zinc chromate at the calcination temperature of 250 °C. At higher calcination temperatures spinel reflections can be found in the XRD patterns. The differences in the XRD patterns of 65Zn35Cr and 33Zn66Cr at the same calcination temperature provide a strong indication that the Zn:Cr ratio plays a crucial role in the phase composition of the calcined catalyst and the morphology of the each component. For catalysts with high Zn:Cr ratio, amorphous ZnO and spinel are formed during calcination. Considering the stoichiometry of the Zn-Cr hydrotalcite-like compound Zn4Cr2(OH)12(CO3) found in the precursor, it is assumed that for catalysts with Zn:Cr ratios between 100:0 and 66:33, the main phases existing in the precursor are Zn-Cr hydroxycarbonates and Zn(OH)2 or ZnCO3.53 During calcination hydrotalcite-like ZnCr hydroxycarbonate decomposes to form Zn-Cr spinel and ZnO, which is also the product of the decomposition of Zn(OH)2 and ZnCO3. For catalysts with Zn:Cr ratios smaller than 66:33, a considerable amount of Cr(OH)3 is also present in the precursor. Hence, ZnCrO4 was formed by oxidation of Cr2O3 in the present of ZnO and O2. At high temperatures ZnCrO4 can decompose back to ZnO and spinel with a relatively large particle size. Thus, the higher Cr content the catalyst has, the more ZnCr2O4 is formed through this pathway resulting in a lower surface area and more ZnCrO4 in the catalyst.

Page 6 of 15

phonon scattering processes of ZnO. This observation indicates that for catalysts with Zn.Cr ratios higher than 33:66, part of the additional ZnO is dissolved in ZnCr2O4 forming the non-stoichiometric Zn-Cr spinel. The bands at around 871 cm−1, 890 cm−1 and 950 cm−1 are ascribed to the Cr–O stretching vibration of chromates. Catalysts with low Zn:Cr ratio seem to exhibit more intense bands at these positions. This observation is consistent with the XRD results indicating that the formation of chromates during calcination is more severe leading to larger chromate particle sizes for the Cr-rich catalysts.

Figure 4. Raman spectra of the ZnO/Cr2O3 catalysts with different Zn:Cr ratios.

Figure 3. XRD patterns of the ZnO/Cr2O3 catalysts with Zn:Cr ratio of 65:35 (left) and 33:66 (right) calcined at 250 °C, 320 °C and 500 °C for 3 h.

Figure 4 shows the Raman spectra of the catalysts with different Zn:Cr ratios. The observed Raman modes and data reported in literature are summarized in Table S1. Based on the selection rules, non-defective ZnCr2O4 spinel should exhibit three Raman active modes, namely A1g, Eg, and F2g, which are observed in Figure 4. The sharp peak at 175 cm-1 originates from the Zn–O bending vibration of the ZnCr2O4 spinel. The band with medium intensity located at 499 cm−1 and the less intense bands at 599 cm-1 and 586 cm-1 can be assigned to the Cr–O stretching vibration (F2g). The intense band at 666 cm−1 is attributed to the symmetric Cr–O stretching vibration of A1g symmetry originating from the CrO6 groups in the spinel structure. Only weak and broad peaks are found at 300 cm−1 belonging to the vibration mode associated with the multiple-

The XP spectra of the calcined ZnO/Cr2O3 catalysts with different Zn:Cr ratios are shown in Figure 5. For all the catalysts, the peaks were observed at almost the same position. The Zn 2p3/2 peak at 1021.2 eV and the Zn 2p1/2 peak at 1044.2 eV originate from the Zn2+ species in the near-surface region of the catalysts. It is not possible to determine whether the Zn2+ ions are present in ZnO or ZnCr2O4 due to their similar peak position.54 The Cr 2p3/2 peak at 576.3 eV and the Cr 2p1/2 peak at 585.9 eV indicate the presence of Cr3+ species as in ZnCr2O4.55 The Cr 2p3/2 peak at 579.2 eV confirms the presence of a small amount of surface chromate species resulting from the oxidation during calcination. The atomic Zn:Cr ratio both in the bulk and on the surface of the catalysts are summarized in Table 2. None of the catalysts exhibits significant zinc enrichment in the near-surface region. Instead, a nonsignificant chromium enrichment of the surface is observed. Therefore, a core-shell structure with a thick amorphous or highly dispersed ZnO shell on ZnCr2O4 can be excluded. The Zn-Cr spinel is assumed to be the dominant state of Zn2+ in the near-surface region. This spinel structure is non-stoichiometric due to the dissolution of ZnO as indicated by the Zn:Cr ratios higher than 1:2 derived from the XPS results.

6

ACS Paragon Plus Environment

Page 7 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Catalysis

Table 2. Bulk and surface composition of the calcined catalysts with different Zn:Cr ratios before and after reaction. Sample

Zn:Cr ratio before reaction (bulk)

Zn:Cr ratio before reaction (surface)

Zn:Cr ratio after reaction a (surface)

70Zn30Cr

71:29

63:37

58:42

65Zn35Cr

67:33

59:41

56:44

55Zn45Cr

57:43

53:47

50:50

a

At 60 bar and 260 °C-300 °C for 60 h (20 h for each temperature step) with a space velocity of 4800 ml h−1g−1. A syngas mixture with H2:CO:N2=54:36:10 was used as feed.

Figure S2 shows the H2 TPR profiles of the calcined catalysts with different Zn:Cr ratios. Pure ZnO (100Zn) exhibits no reduction peak, while the ZnO/Cr2O3 catalysts with Zn:Cr ratios from 80:20 to 55:45 show a single well-defined reduction peak at about 310 °C. The single reduction peak of 40Zn60Cr and 33Zn66Cr shifts to 353 °C and 384 °C, respectively. It can be observed that its intensity increases with decreasing Zn:Cr ratio. For 20Zn, three broad reduction peaks located at 457 °C, 512 °C and 670 °C are found. The H2 TPR profile of 100Cr consists of overlapping peaks at 576 °C and 673 °C. In addition, two sharp shoulders are found between 550 °C and 580 °C. In literature,56-58 the single reduction peak for catalysts with Zn:Cr ratios from 80:20 to 40:60 was ascribed to the reduction of the Zn-Cr spinel. Based on the XRD, XPS and Raman results in the present work, this peak more likely originates from the reduction of zinc chromates. For 40Zn60Cr and 33Zn66Cr, the shift of this reduction peak to the high-temperature region is caused by the larger chromate particle size. The complicated peaks at high temperature for 20Zn80Cr and 100 Cr originate from the reduction of sodium chromates, which can be further confirmed by the XRD results of these samples after the TPR experiments (Figure S3). The observed reflections of sodium chromite, which is the reduction product of sodium chromate, prove the existence of sodium chromate. Furthermore, the total H2 consumption (Table S2) increases with decreasing Zn:Cr ratio indicating that the catalysts with high levels of Cr contain a higher amount of chromates species.

Figure 5: Zn 2p (left) and Cr 2p (right) XP spectra of the calcined ZnO/Cr2O3catalysts with different Zn:Cr ratios.

Figure 6. Methanol productivity of catalysts with different Zn:Cr ratios at 260-300 °C, 60 bar H2:CO:N2=54:36:10 and a flow of 4800 ml h−1g−1.

7

ACS Paragon Plus Environment

ACS Catalysis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

b. Catalytic tests The effect of the Zn:Cr ratio on the performance of the catalysts was studied in methanol synthesis. The carbon balance for all the catalytic tests was between 99% and 101%. At all three reaction temperatures the highest methanol productivity was achieved with 65Zn35Cr (Figure 6). The selectivities to methanol and the undesired product hydrocarbons as well as the degrees of CO conversion are summarized in Table 3. For all the catalysts, CO conversion increases with increasing reaction temperature, but the rate of hydrocarbon formation increases more rapidly with rising reaction temperature than that of methanol formation resulting in a lower methanol selectivity at 300 °C. The formation of hydrocarbons over ZnO/Cr2O3 catalysts has also been observed by Jiao et al.14 When the reaction temperature was high enough (400 °C), ZnCr oxide catalysts prepared by co-precipitation exhibited almost 100% selectivity towards hydrocarbons and CO2 combined. In contrast to the 100Cr catalyst without Zn, the 100Zn catalyst without Cr achieves a significant methanol productivity. The catalysts with Zn:Cr ratios in the range from 70:30 to 50:50 reach high degrees of conversion, high methanol selectivity and low selectivity to hydrocarbons in comparison to the catalysts containing an excess of Cr. Generally, a high Cr content leads to high acidity of the catalyst, which favors the formation of hydrocarbons. Water as the coupled product of hydrocarbon formation further enhances the formation of CO2. As a result of the enhanced formation of both hydrocarbons and CO2, the selectivity to methanol significantly decreases. Furthermore, the high content of Na residue may also play an important role in the low conversion and high selectivity toward hydrocarbons for the catalysts containing excess of Cr. As shown in Table 1, with decreasing Zn:Cr ratio from 50:50 to 0:100 the content of the Na residue increases drastically leading to the formation of more chromate species, which further results in the rapid decrease in surface area. As a result, the degree of conversion for these catalysts decreases with increasing Na content. In addition, the selectivity towards high products also significantly increases with increasing Na content. This observation corresponds to the promoting effect of alkali metals on the formation of long-chain products, which has been reported in both Fischer-Tropsch synthesis59-60and higher alcohol synthesis.54-55 Additionally, a series of Na-free catalysts was synthesized by using (NH4)2CO3 as the precipitation reagent, while other preparation and test conditions were not changed. As shown in Figure S4, the Na residue was confirmed to cause the drastic decrease in surface area. Correspondingly, for catalysts with Zn:Cr